High reliability computer systems and telecommunications systems are fully duplicated. Each of these systems has two subsystems each capable of performing the system operations. One subsystem is active performing the operations and is commonly called the active subsystem. The other subsystem is waiting to take over the system operations should the active subsystem fail and is commonly) referred to as the standby system. Unfortunately, the standby does not have all of the dynamic data necessary to immediately take ove operations of the system. This data resides on printed circuit boards within the active system. During an unplanned switch from the active subsystem to the standby subsystem, the system operations will not be performed for a period of time. High reliability computeps and telecommunications systems are measured by the projected unavailability. This unavailability is normally measured in minutes per years. An unplanned transfer from the active system to the standby system makes the system unavailable for a period of time and increases the projected unavailability time of the system. The necessity to transfer from the active subsystem to the standby subsystem normally results from a printed circuit board failure within the active system or a programming error. However, another cause is the inadvertent removal of circuit boards from the active subsystem.
The proper procedure to be followed by field technicians when removing a printed circuit board from the active system is to cause the transfer from the active subsystem to the standby subsystem under control of a computer controlling the system. When this computer is aware that the transfer is going to take place, the computer executes procedures such that the system is continuously available during the transfer of control from the active subsystem to the standby subsystem. Unfortunately, field technicians do fail to follow the proper procedures and mistakely remove circuit boards from active subsystems.
U.S. Pat. No. 5,196,712 provides a partial solution for detecting the removal of a circuit board. However, the apparatus disclosed in that U.S. patent is implemented using a mechanical switch which requires that the printed circuit board's latch handle have a plate that inserts into an optical housing mounted on the circuit board. The disclosed technique does not lend itself to modem circuit board manufacturing techniques such as surface mount technology. The housing of necessity must protrude a significant distance above the surface of the circuit board causing two problems. The first problem is that it lowers the number of printed circuit boards that can be inserted into a printed circuit board carrier. The second problem is that because of the bulk of the housing, it cannot be mounted on the printed circuit board using surface mount technology. Another problem with the disclosed technique is that it is vulnerable to false indications that the circuit board is being removed. These false indications are due to the fact that the proposed technique utilizes circuitry in which a single failure can falsely indicate that the circuit board is being removed.
Another problem with the method set forth in U.S. Pat. No. 5,196,712 is that the rear portion of a circuit board is densely populated with a variety of integrated circuits, test jacks, and LED indicators. The above referenced patent compounds this problem since it utilizes a great deal of circuit board area. In addition, LED indicators, where the LED must be mounted in a plane horizontal to the circuit board, cannot be manufactured using surface mount technology. In addition, mounting LEDs in this manner results in lower reliability than if the LEDs are mounted utilizing surface mount technology.